--- trunk/Split.pm	2004/08/08 19:22:56	4
+++ trunk/Split.pm	2004/12/17 18:32:34	7
@@ -4,7 +4,7 @@
 use strict;
 use warnings;
 
-our $VERSION = '0.00';
+our $VERSION = '0.01';
 
 use SWISH::API;
 use Text::Iconv;
@@ -17,6 +17,11 @@
 
 use Data::Dumper;
 
+use constant {
+	ADDED => 1,
+	DELETED => 2,
+};
+
 =head1 NAME
 
 SWISH::Split - Perl interface to split index variant of Swish-e
@@ -37,7 +42,7 @@
 engines) easier.
 
 In the background, it will fork swish-e binaries (one for each index slice)
-and produce UTF-8 encoded XML files for it. So, if your imput charset isn't
+and produce UTF-8 encoded XML files for it. So, if your input charset isn't
 C<ISO-8859-1> you will have to specify it.
 
 =head1 Methods used for indexing
@@ -144,14 +149,14 @@
 	my $swishpath = shift || return;
 	my $data = shift || return;
 
-	my ($out,$err) = $self->put_slice($swishpath, $self->to_xml($data));
+	my $slice = $self->put_slice($swishpath, $self->to_xml($data));
 
-	if ($err) {
-		carp "$swishpath: $err";
-		return 0;
-	}
+#	if ($err) {
+#		carp "$swishpath: $err";
+#		return undef;
+#	}
 
-	return 1;
+	return $slice;
 }
 
 =head2 delete
@@ -167,29 +172,37 @@
 
 	my @paths = @_ || return;
 
+	foreach my $path (@paths) {
+		$self->{'paths'}->{$path} = DELETED; 
+	}
+
 	return 42;
 }
 
 
-=head2 finish
+=head2 done
 
 Finish indexing and close index file(s).
 
-  $i->finish;
+  $i->done;
 
 This is most time-consuming operation. When it's called, it will re-index
 all entries which haven't changed in all slices.
 
 Returns number of slices updated.
 
+This method should really be called close or finish, but both of those are
+allready used.
+
 =cut
 
-sub finish {
+sub done {
 	my $self = shift;
 
 	my $ret = 0;
 
 	foreach my $s (keys %{$self->{'slice'}}) {
+		print STDERR "closing slice $s\n";
 		$ret += $self->close_slice($s);
 	}
 
@@ -212,6 +225,11 @@
 
 sub swishpaths {
 	my $self = shift;
+
+	my $s = shift || return;
+	return if (! exists($self->{'slice'}->{'s'}));
+
+	return keys %{$self->{'slice'}->{'s'}};
 }
 
 =head2 swishpaths_updated
@@ -313,7 +331,6 @@
 sub find_paths {
 	my $self = shift;
 
-	my $s = shift || return;
 }
 
 
@@ -330,9 +347,7 @@
 If you want to see what is allready defined for swish-e in configuration
 take a look at source code for C<DEFAULT_SWISH_CONF>.
 
-It uses C<cat> utility to comunicate with C<swish-e>. Path is provided
-by C<File::Which>. Do Windows users have to change that to C<COPY /B>
-or something similar?
+It uses C<stdin> as C<IndexDir> to comunicate with C<swish-e>.
 
 =cut
 
@@ -351,8 +366,7 @@
 	print $tmp_fh <<"DEFAULT_SWISH_CONF";
 # swish-e config file
 
-IndexDir cat
-#SwishProgParameters -
+IndexDir stdin
 
 # input file definition
 DefaultContents XML*
@@ -410,7 +424,7 @@
 
 	print STDERR "creating slice $s\n";	# FIXME
 
-	my @swish = qw(swish-e -S prog -c);
+	my @swish = qw(swish-e -u -S prog -c);
 	push @swish, $swish_config;
 
 	## Build the harness, open all pipes, and launch the subprocesses
@@ -430,9 +444,11 @@
 
 =head2 put_slice
 
-Pass XML data to swish and receive output and errors.
+Pass XML data to swish.
 
-  my ($out,$err) = $i->put_slice('/swish/path', '<xml>data</xml>');
+  my $slice = $i->put_slice('/swish/path', '<xml>data</xml>');
+
+Returns slice in which XML ended up.
 
 =cut
 
@@ -456,6 +472,7 @@
 	$self->{'slice'}->{$s}->{'in'} .=
 		"Path-Name: $path\n".
 		"Content-Length: ".(length($xml)+1)."\n".
+		"Update-Mode: Index\n".
 		"Document-Type: XML\n\n$xml\n";
 
 	# do I/O
@@ -463,6 +480,8 @@
 
 	$self->slice_output($s);
 
+	$self->{'paths'}->{$path} = ADDED; 
+
 	return $s;
 }
 
@@ -470,7 +489,7 @@
 
 Prints to STDERR output and errors from C<swish-e>.
 
-  $i->slice_output($s);
+  my $slice = $i->slice_output($s);
 
 Normally, you don't need to call it.
 
@@ -488,18 +507,18 @@
 	if (length $self->{'slice'}->{$s}->{'out'} > $self->{'slice'}->{$s}->{'out_len'}) {
 		#print STDERR "swish-e OUT: ",$self->{'slice'}->{$s}->{'out'},"\n" if ($self->{'slice'}->{$s}->{'out'});
 		$self->{'slice'}->{$s}->{'out_len'} = length $self->{'slice'}->{$s}->{'out'};
-		return 1;
+		return $s;
 	} elsif (length $self->{'slice'}->{$s}->{'err'} > $self->{'slice'}->{$s}->{'err_len'}) {
 		print STDERR "swish-e ERR: ",$self->{'slice'}->{$s}->{'err'},"\n" if ($self->{'slice'}->{$s}->{'err'});
 		$self->{'slice'}->{$s}->{'err_len'} = length $self->{'slice'}->{$s}->{'err'};
 		# this is fatal
-		return 0;
+		return undef;
 	}
 
-	return 1;
+	return $s;
 }
 
-=head2 close_slice {
+=head2 close_slice
 
 Close slice (terminates swish-e process for that slice).
 
@@ -523,7 +542,7 @@
 	$self->slice_output($s);
 
 	# clean up
-	$self->{'slice'}->{$s}->{'h'}->finish or confess "finish on slice $s returned: $?";
+	$self->{'slice'}->{$s}->{'h'}->finish or confess "finish on slice $s returned $?: $! -- ",$self->{'slice'}->{$s}->{'err'};
 	
 	delete($self->{'slice'}->{$s}) && return 1;
 	return 0;
@@ -567,7 +586,7 @@
 __END__
 
 
-=head2 Searching
+=head1 Searching
 
 Searching is still conducted using L<SWISH::API>, but you have to glob
 index names.
@@ -582,11 +601,11 @@
 That would also benefit performance, but it increases indexing time
 because merged indexes must be re-created on each indexing run.
 
-=head2 EXPORT
+=head1 EXPORT
 
 Nothing by default.
 
-=head2 EXAMPLES
+=head1 EXAMPLES
 
 Test script for this module uses all parts of API. It's also nice example
 how to use C<SWISH::Split>.